31 March 2008 Carlo Comporti CESR Secretary General The Committee of European Securities Regulators 11-13 avenue de Friedland 75008 Paris ESF/SIFMA RESPONSE TO CESR CONSULTATION PAPER ON THE ROLE OF CREDIT RATING AGENCIES IN STRUCTURED FINANCE Dear Mr Comporti The European Securitisation Forum 1 (ESF) and Securities Industry and Financial Markets Association 2 (SIFMA) are pleased to respond to the February 2008 CESR Consultation Paper (CP) on the role of Credit Rating Agencies (CRA) in Structured Finance (SF). This response reflects preliminary feedback from ESF and SIFMA members in Europe. Please note that as a global organisation, SIFMA is likely to enter into a broader study of the role of CRAs. The results of that study could change the emphasis of our initial recommendations, which we are providing now so as to not impede the CESR process. We will forward a copy of this broader study reflecting the position of SIFMA members globally to CESR when it is completed. In this context we note the recent initiative of the UK and US governments to set up a joint working group of regulators to monitor and regulate the banking system, inter alia examining the role of CRAs in evaluating risk. We believe that CRAs play a very important role in the capital markets, particularly in SF where they have been critical to increasing the availability of credit and distribution of risk. Nevertheless the recent market turmoil and perceived weaknesses in ratings have given rise to legitimate concerns, including perceived conflicts of interest, the transparency of ratings and underlying methodologies, and the suitability of the current self-regulatory regime. Please find attached in Annex I, our response to the 18 questions of the CP. We would be pleased to discuss our response or any further issues with CESR at your convenience. Yours sincerely 1 The ESF is the voice of the securitisation and CDO marketplace in Europe, with the purpose of promoting efficient growth and continued development of securitisation throughout Europe. Its membership is comprised of over 150 institutions involved with all aspects of the securitisation and CDO business, including issuers, investors, arrangers, rating agencies, legal and accounting advisors, stock exchanges, trustees, IT service providers and others. The ESF is affiliated with SIFMA. The ESF has two sister organisations: the American Securitization Forum and the Asia Pacific Securitisation Association. 2 SIFMA brings together the shared interests of more than 650 securities firms, banks and asset managers. SIFMA s mission is to promote policies and practices that work to expand and perfect markets, foster the development of new products and services and create efficiencies for member firms, while preserving an enhancing the public s trust and confidence in the markets and the industry. SIFMA works to represent its members interests locally and globally. It has offices in London, New York, Washington DC, and its sister Association, the Asia Securities Industry and Financial Markets Association (ASIFMA), is based in Hong Kong. London New York Washington Hong Kong St. Michael s House 1 George Yard London EC3V 9DH P: 44.20.77.43.93.00 F: 44.20.77.43.93.01 www.sifma.org
Rick Watson Managing Director European Securitisation Forum rwatson@europeansecuritisation.com Bertrand Huet Delarherse MD, European Legal & Regulatory Counsel SIFMA
ANNEX I: ESF/SIFMA RESPONSE TO CESR CONSULTATION PAPER ON THE ROLE OF CREDIT RATING AGENCIES IN STRUCTURED FINANCE TRANSPARENCY OF RATINGS PROCESSES AND METHODOLOGIES Q1: Do you agree that CRAs need to make greater on-going efforts to clarify the limitations of their ratings? We broadly agree. Although there have recently been improvements in the level of communication with the market on this issue, CRAs can and should do more to communicate on an ongoing basis that their SF ratings comprise only a snapshot assessment of the probability of default and/or expected loss (depending on the specific CRA) and are not indicators of stability, liquidity and volatility. In this context, we would not support a new rating scale for SF products, nor do we support a modification of the existing scale or the addition of suffixes. We would instead recommend the separate provision of additional analysis of deal structures and of risk characteristics, as well as the disclosure of the fundamental criteria which underlie the models used to rate securities as the most effective way of improving transparency. There is some question as to the appropriateness of providing a scoring on other risk factors but we generally believe that trying to finitely score specific risk factors would be confusing to market participants. We also believe that the practical difficulties of implementing such changes to the ratings scale (including e.g. implementing changes to the myriad of references to the current ratings scale in numerous investment guidelines and laws etc) would be significant. We also note that the purpose of SF is fundamentally to repackage credit with assets of similar characteristics and not to minimize the level of credit protection needed (see paragraph 40 of the CESR CP). Q2: Do you agree that although there has been improvement in transparency of methodologies, the accessibility and content of this information for complex SF products requires further improvement in particular so that investors have the information needed for them to judge the impact of market disruption on the volatility of the ratings? We broadly agree. While there have been improvements in CRA disclosure of methodologies we would welcome further enhancements in the accessibility and content of this information. CRAs can and should make efforts to improve the navigability of their web-sites and make it easier to track changes to methodologies. CRA staff should also be available at short notice to discuss changes to methodologies or underlying assumptions with investors and other market participants In terms of content, we would welcome more information on the key assumptions and correlations underpinning the methodologies. We recommend that CRAs disclose the specific assumptions used in modeling and rating transactions. Greater disclosure of information such as the base case probability of default or expected losses of the underlying assets of a deal (and other key figures) when permitted under data protection laws and consistent with CRA confidentiality obligations under the IOSCO Code of Conduct would provide investors and other market participants more insights on the transactions. Q3: Do you agree that there needs to be greater transparency regarding the specific methodology used to determine individual SF ratings as well as rating reviews? We agree. Whilst methodologies (at least for CDO transactions) are freely available it is sometimes difficult to track which particular methodology has been used to rate a specific issue and whether a rating review has been triggered by a change in that methodology or the performance of the underlying asset pool. To ensure a level playing field across asset classes, we agree that CRAs should highlight clearly to investors which particular methodology a SF rating is based on, together with clear indications of the specific location where these can be found. CRAs should also be clear as to whether a ratings review has been triggered by a change in methodology or a change in the performance of the underlying asset pools. In this context we would note that methodologies and underlying assumptions/correlations might change as a result of better insights gained from the performance of the underlying asset pools. In the interests of greater transparency, we would also suggest that CRAs disclose their policies on notching. Q4: Do you agree that there needs to be greater public and standardised information on structured products in the EU? How would this best be achieved?
We entirely agree. The ESF has initiated and continues to lead a series of industry work-streams designed to ensure that the timeliness, frequency, quality, accessibility and uniformity of information for rated SF products continue to improve. We have e.g. recommended that SF issuers use more standardized reporting for pool performance to facilitate the valuation process for investors and other market participants. We have thus published the Securitisation Market Practice Guidelines to spear head efforts to improve and standardise reporting practices and definitions across Europe. MONITORING OF RATINGS PERFORMANCE Q5: Do you agree that contractually set public announcements on SF performance would not add sufficient value to the market to justify the cost and possible saturation of the market with non-material information? We agree. In light of the automatic monitoring already carried out on SF products, the wholesale nature of the SF market and large volume of data already published, we do not believe that imposing contractually determined public reviews would add value to the monitoring process. On the contrary, there would likely be high costs associated with such measures and the large increase in the volume of CRA published information would likely have a negative impact on market transparency as a whole. Q6: Do you agree that the monitoring of SF products presents significant challenges, and therefore should be a specific area of oversight going forward? Are there any particular steps that CRAs should take to ensure the timely monitoring of complex transactions? We broadly agree. Monitoring is important and we support the steps taken so far by CRAs to increase resources dedicated to the surveillance of SF transactions. We would further encourage CRAs to build on these steps and consider greater flexibility in shifting more resources to surveillance when appropriate. We note that this resource (re)allocation should be proportionate to the fundamental causes of weaknesses in ratings. We do not consider precipitous downgrades to be primarily a function of suboptimal monitoring processes but rather due to the rapid increase in the deteriorating performance of the underlying assets coupled with shortcomings in the model applied at the time of the initial rating. CRA STAFF REOURCING As an industry, our overriding interest is in the quality and completeness and timeliness of the ratings produced rather than in the quality, quantity and costs of the human resources required by CRAs to produce this output. Our response to questions 7-10 should be read in that light. Q7: Do you believe that CRAs have maintained sufficient human resource, both in terms of quality and quantity, to adequately deal with the volumes of business they have been carrying out, particularly with respect to SF business We broadly agree, although there is room for improvement. The availability of technically qualified and experienced staff for the most complex products is an issue not only for CRAs but also for other financial institutions (including issuers, arrangers and investors). In the CRA context, we note that agencies have made great efforts to educate their own staff (as well as investors) and the while CRA analysts do not generally spend their entire working lives with the same employer, rating methodologies and other supporting tools are in writing and can be passed from one analyst to another. Q8: Do you consider that the generally unaltered educational and professional requirements of CRAs' recruitment policies negatively impact the quality of their rating process, given the rising complexity of SF products? We disagree. We do not consider that there is a need to impose CRA-specific educational and/or professional requirements. This would not add significantly to the quality of ratings. Experience, tenure and range of knowledge are all factors that are as important as academic qualifications. Q9: Do you agree there is a need for greater transparency in terms of CRA resourcing?
We note that the CRAs were not able to provide information to CESR to allow a more detailed analysis of their staffing trends. We agree that CRAs should have sufficiently advanced human resource management processes to be able to provide key trends in staffing, employee development and turnover levels. Q10: Do you agree that more clarity and greater independence is required for analyst remuneration at the CRAs? We accept that additional (anonymised) information on staff turnover at different levels within the CRA and greater clarity on the remuneration policies (as opposed to actual salaries) of individual SF analysts may be useful in conflict-of-interest management terms. However, we are very much aware that such increases in transparency may reduce CRA competition given the oligopolistic structure of the industry. CONFLICTS OF INTEREST Q11: Do you see the level of interaction between CRAs and issuers of SF products creating additional conflicts of interest for the CRAs to those outlined above? Do you believe that any of these conflicts are not being managed properly? We do not view the interaction between CRAs and SF issuers as giving rise to additional conflicts (cf. the interaction between CRAs and corporate issuers) and believe that the conflicts which do arise are currently managed appropriately. Nevertheless we acknowledge the nature of the rating process for SF products may give rise to the perception of greater conflict risk which must be managed properly. The heterogeneity and complexity of SF deals as well as the focus on the credit worthiness of the underlying assets rather than the issuer means that CRA will engage in more frequent discussions with issuers as they need to ensure that they fully understand the proposed structures. We believe that the perceived potential conflicts associated with this interaction need to be managed through detailed and rigorous policies and procedures. For example, we fully support the IOSCO Code recommendations for CRAs to review the past work of analysts that leave the employ of the CA and join an issuer the CRA rates and for the CRA to review employee remuneration policies to ensure that these do not compromise the CRA s rating process. We further note the powerful commercial mitigant to conflict in both the SF as well as the corporate areas is the potential reputational risk to the CRAs if rating integrity is compromised. Q12: Do you agree that greater transparency is required regarding the nature of interaction between CRAs and issuers/arrangers with regards to SF products and that there needs to be clearer definitions of acceptable practice? We broadly agree. CRAs should ensure they are fully transparent about their interaction with issuers/arrangers of SF products and that they have strong policies and procedures to monitor and control this interaction and ensure it reflects their public position. Q13: Do you believe there needs to be greater disclosure by CRAs over what they consider to be ancillary and core rating business? CRAs should maintain the current level of separation between SF credit analysts and commercial staff. Certain CRAs provide ex poste ancillary services and advisory services. In certain cases research is provided as part of the ratings process during a transaction while at other CRAs there is a fee for such services. These research services generally represent a fraction of the revenue generated by the ratings/analytical business. As in the corporate world and partly as a result of the IOSCO Code requirement (for CRAs to separate their credit rating business and analysts from any other business that may present a conflict of interest) all major CRAs have gone to great lengths to ensure that adequate separation exists between SF credit analysts and commercial staff. This separation also goes to a more senior management level than used to be the case. Q14: Do you believe that the fee model used for SF products creates a conflict of interest for the CRAs? If yes, is this conflict of interest being managed appropriately by the CRAs? We disagree but acknowledge that the payment of fees by issuers could create a perception of conflict unless strong mitigants are put in place. Ratings fees are generally paid by the SPV issuers. In theory (as with all other ratings products where fees are paid by issuers) this could create a perception that a conflict of interest could arise. This conflict analysis is no different than with a corporate rating and CRAs have established policies and procedures to manage such
conflicts. We also believe that the IOSCO Code (which has been implemented by CRAs covering a large proportion of the SF rating industry) adequately addresses this concern. In addition, we believe that the most powerful mitigant to conflicts in both the SF and corporate area is the potential reputational risk to the CRA if its ratings integrity is compromised. Q15: Do you agree that there needs to be greater disclosure of fee structures and practices with regard to SF ratings so as to mitigate potential conflicts of interests? Whilst we would welcome greater transparency on fee disclosure, we believe this is a commercial issue best left to the relevant parties to agree upon. Although on many products each CRA has a standard fee scale that is openly disclosed, in many circumstances the actual fee will reflect the characteristics of the transaction including its structure, complexity and innovation compared with similar deals. This variability of fees in the SF industry is a function of the amount of CRA staff time required to analyze the transaction as well as deal size complexity and whether the issuer is a first time or frequent/repeat issuer. Public disclosure surrounding 'break-up' fees, when an arranger is charged for analytical work done by a CRA even if it chooses not to proceed with the rating, is not always made by CRAs. Requiring CRAs to disclose break-up fees carries the risk that this practice might motivate arrangers to request ratings from less CRAs to reduce possible negative publicity. Consideration should be given as to whether it should be disclosed as to whether a CRA has been approached for a rating and then had the request for rating withdrawn. This issue is also being considered by IOSCO in its consultation on the CRA Code of Conduct, and the industry plans to address this question in that response. REGULATORY OPTIONS Q16: Do you agree with CESR's view of the benefits and costs of the current regime? We broadly agree with the CESR cost / benefit analysis of the current (self-regulatory) regime. In this context, we would emphasize the significant benefits of the current regime not least the proven ability of market forces to ensure that CRA address past shortcomings and deliver a product with the integrity, timeliness and completeness required by the market. On the cost side and referring to the perceived weakness of the incentive for CRAs to comply with a self-regulatory regime, we would not want to underestimate the reputational risk associated with non-compliance with the IOSCO Code. Q17: Do you agree that CESR has correctly identified the likely benefits and costs related to formal regulatory action? We broadly agree that CESR has correctly identified the benefits and costs of introducing a formal regulatory regime. However, we would note that the global nature of the rating business means that the costs and benefits of formal regulation should be analysed in a global context. The analysis should moreover take account of any existing regimes such as that pertaining to External Credit Assessment Institutions in the EU. With respect to the concrete cost/benefit analysis we would highlight in particular the danger of over-reliance on CRAs and de-emphasize the incentive of regulatory compliance compared to that associated with reputational risk. Q18: Do you believe that the current self-regulatory regime for CRAs should be maintained rather than introducing some form of formal recognition/regulation? More regulation will likely lead to even greater inappropriate reliance on ratings and micro regulations that could adversely impact the opinion concept that underlies current rating. Imposing a formal recognition/regulatory regime for CRAs will likely make CRAs more bureaucratic and less responsive to changing market conditions which would have the effect of reducing their relevance.